DE69413504T2 - DEVICE AND METHOD FOR DOSE PLANNING IN RADIATION THERAPY - Google Patents

Description

This invention relates generally to a radiotherapy system for the treatment of tumors or the like, and more particularly to a radiotherapy planning method for calculating the necessary beam intensities in a therapy system for delivering multiple beams at different angles in a plane to precisely regulate the radiation dose in irregularly shaped zones in the patient.

A medical radiotherapy facility treats tumor tissue with high-energy radiation. To ensure that the tumor receives enough radiation to be destroyed, as well as to minimize damage to surrounding and adjacent non-tumor tissue, the dose and location of the dose must be precisely controlled.

In internal source radiotherapy, capsules of radioactive material are placed near the tumor tissue in the patient. The dose and location are precisely controlled by the physical positioning of the isotope. However, internal source radiotherapy has the disadvantages of any surgical procedure, including patient discomfort and the risk of infection.

External source radiotherapy uses a radiation source that is external to the patient, typically either a radioisotope such as 60Co or a high-energy X-ray source such as a linear accelerator. The external source produces a beam of radiation directed into the patient toward the tumor site. External source radiotherapy avoids some of the problems of internal source radiotherapy, but undesirably and necessarily irradiates a significant volume of non-tumor or healthy tissue in the path of the radiation beam along with the tumor tissue.

By projecting the external radiation beam into the patient at different "stage" angles, with the beams converging at the tumor site, the adverse effect of irradiating healthy tissue can be reduced while maintaining a given radiation dose in the tumor tissue. The particular volume elements of healthy tissue along the path of the radiation beam are changed, reducing the total dose for each such element of healthy tissue during the overall treatment.

The irradiation of healthy tissue can also be reduced by narrowly focusing the radiation beam on the general cross-section of the tumor taken perpendicular to the axis of the radiation beam. Numerous systems exist for creating such circumferential focusing, some of which use multiple sliding closures that can piecemeal create a radiopaque mask of any shape.

To allow the treated area to be located at a location other than the center of rotation, the offset angle of the radiation beam with respect to a radius line between the radiation source and the center of rotation of the radiation source can be adjusted to the outline of the tumor as part of the focusing of the beam. At the same time, changing the offset angle and the width of the radiation beam as a function of the stage angle allows the tumor tissue to be treated in a plane parallel to the irradiation beam aimed precisely at the target may have an irregular cross-section. The width and offset angle of the irradiation beam can be controlled by using a folding collimator.

Adjusting the offset angle, center point and size of the radiation beam at different stage angles allows considerable latitude in controlling the dose. Particularly when the tumor is concave or highly irregular, these approaches still result in a significant amount of unwanted dose being transferred to healthy tissue.

In copending U.S. patent application Serial No. 07/865,521, filed March 19, 1992 and assigned to the same assignee as the present invention, a radiation therapy machine is described which provides greatly reduced radiation exposure to healthy tissue. In this radiation therapy machine, a number of radiation attenuation blades are mounted in a frame in the radiation beam prior to the beam entering the patient. To block a given single beam of the beam, the blades are inserted into the radiation beam in a closed state, and to allow unobstructed passage of a given single beam of the beam, the blades are inserted out of the radiation beam in an open state. By controlling the ratio of time spent in the open and closed states, each individual beam can be attenuated over a continuous range of intensity.

This ability to determine not only the outline of the radiation, but also the intensity of each individual beam enables extremely precise control of the irradiation volume.

Theoretically, with the correct modulation of each individual beam of the beam as it travels through a series of angles around the patient, the radiotherapy machine can precisely target the dose even in concave or irregular zones. It is also important that the radiation can be precisely excluded from zones containing radiation-sensitive organs or the like.

Such dose mapping methods require mapping the dose received by any volume element of the patient among the many possible beams of radiation at different stage angles. This mapping, manifested as a set of beam intensities for each beam and for each stage angle of the beam, is referred to as a treatment sinogram.

The copending application referenced above describes a method of generating a treatment sinogram using techniques similar to the image reconstruction techniques of computed tomography ("CT") with the important difference that CT image reconstruction derives an image of the patient by measuring the attenuations of several individual beams through the patient, while the radiation therapy system imprints a "dose image" on the patient by pre-attenuating a set of radiation beams at the various angles around the patient.

The above-mentioned sinogram generation methods pose a challenge to the computational capabilities of current computer hardware.

WO-A-91/18552 discloses a radiotherapy device provided with a radiation source for generating a radiation beam directed at a patient at a plurality of in-plane stage angles. Each beam of the beam has an independently adjustable particle flux, with a beam following a path at a respective stage angle. Computer simulation techniques are then used to determine an optimal dose efficiency and the beams are controlled based on the optimal dose efficiency.

In order to shield critical organs not affected by the tumor, EP-A-0556874 (which is a document pursuant to Art. 54/3 EPC) discloses a radiotherapy device on which the individual beams are adjustable to obtain irregular irradiation fields. The desired shape of the field is calculated, with sheets of opaque material being arranged to give the desired radiation field.

According to a first aspect of the present invention, there is provided a radiation therapy device comprising a radiation source for generating a radiation beam directed at a volume of a patient at a plurality of stage angles located in a stage plane and comprising a plurality of individual beams distributed in the stage plane, each individual beam having an independently adjustable particle flux, an individual beam following a path at a single stage angle, and a therapy planning device for determining the particle flux of each beam, the therapy planning device comprising:

an electronic computer which in turn contains:

means for receiving a target dose map for a target irradiation dose along a slice through the patient's volume;

wherein the therapy planning device further contains:

a charge calculation device for calculating a distribution of electrical charges in the disk, which would generate a potential energy field corresponding to the target characteristic field; and

an integration device for integrating the charges of the distribution lying along the path to determine the particle flux of the single beam at the stage angle.

According to a second aspect of the present invention, there is provided a method of radiotherapy planning for a radiotherapy device having a radiation source for generating a radiation beam directed at a patient's slice at a plurality of stage angles located in a stage plane and comprising a plurality of individual beams distributed in the stage plane, each individual beam having an independently adjustable particle flux, an individual beam following a path through the slice at a stage angle, the method comprising:

(a) entering a target dose map along a slice through the patient's volume as set digital values into an electronic computer (51), the digital values representing a dose of ionizing radiation necessary for treating the patient;

characterized in that the method further comprises :

(b) calculating a distribution of electrical charges in the disk that would produce a potential energy field corresponding to the target dose map; and

(c) Integrate the charges of the distribution lying along the path to determine the particle flux of the single beam at the stage angle.

Summary of the invention

The present invention thus provides a new method for calculating a treatment sinogram. The method models the target dose distribution in the patient as an electric potential field and uses the techniques of electrostatics to allocate the dose contributed by individual beams at different angles. This method can provide improved computational efficiency for the generation of a treatment sinogram for radiotherapy and is suitable for execution in certain steps on a new and extremely fast to write electrostatic computer.

More specifically, an electronic computer receives data indicating the target dose to be delivered by the multiple beams of the radiotherapy machine along the disk through the patient's volume. A charge calculator receiving this target dose calculates the distribution of electrical charges in the disk that would produce a potential energy field corresponding to the target dose map. To determine the intensity of the beam along that path, an integrator simply integrates these charges along any path through the disk.

One aim of the invention is therefore to provide a simple and rapid method of generating a treatment sinogram from a target dose map. Formulating the problem as an electrostatic problem allows a wide variety of numerical methods to be used in the actual calculation, potentially providing improved performance over approaches based on computed tomography concepts. In one embodiment, the charge calculator may be a conductive foil formed from a plurality of interconnectable conductive elements which are first connected together in at least one dose region to receive a charge and then separated to measure the charge on each isolated element. The integrator is a digital computer for receiving the charge values in a matrix and performing the required interpolation and integration.

Another object of the invention is thus to provide a method of generating a treatment sinogram that is easily adapted to a high speed analog computer device. The conductive foil calculates the charge distribution using actual charges on a substantially instantaneous basis. Each charge calculates its own location simultaneously with other charges in a manner similar to, but significantly faster than, a massively parallel digital computer.

The target dose map may have at least two zones: a first tumor zone that is to receive at least a treatment radiation threshold and a second organ zone that is to receive at most a tolerance radiation threshold. The charge calculator calculates the charge that the potential energy field corresponding to an instantaneous dose map (which is initially equal to the target dose map) in each zone would generate while the integrator is moving along a given path through the Disc integrated to determine the individual beam weight along that path.

To produce a sinogram of flux values, by repeating the integration of the charges, means are provided for determining the weight for many individual beams around different angles of the stage, means being provided for zeroing any negative flux values in the sinogram to produce an adjusted sinogram. A dose calculator, receiving this adjusted sinogram, determines the actual dose that would be produced by this adjusted sinogram. The instantaneous dose map is modified according to the difference between the target dose and the actual dose. In one embodiment, the instantaneous dose map is modified for the next iteration by subtracting the difference from the instantaneous dose map.

Another object of the invention is therefore to allow maximum dose limitations to be assigned to certain zones to protect radiation sensitive organs. These imposed limitations may produce an initial sinogram with negative single beam weights, a condition that is not physically feasible. However, it has been found that an iterative correction of the instantaneous dose map based on a calculation of the actual dose with negative single beam weights truncated to zero can be used to produce a sinogram that converges to the target dose.

The above and other objects and advantages of the invention will become apparent from the following description. In the description, reference is made to the accompanying drawings which form a part hereof, and in which For illustration purposes, several expedient embodiments of the invention are shown.

Fig. 1 is a perspective view of the compensator assembly used in a radiotherapy machine that can be used to practice the present invention, showing the compensator blades and their associated actuators;

Fig. 2 is a cross-section of the compensator assembly of Fig. 1 taken along line 2-2, showing, for a fan beam, the trapezoidal appearance of each compensator blade and the guide rails for supporting the compensator blades during their translation;

Fig. 3 is a perspective cut-away view of a set of guide rails and a blade of Fig. 2, showing a collar for supporting the blade in its fully closed position;

Fig. 4 is a block diagram showing the elements of a radiotherapy apparatus including a conventional CT scanner and the compensator and including a computer suitable for controlling that compensator;

Fig. 5 is a simplified elevation of a cross-section of the patient showing a tumor and an adjacent radiosensitive organ with multiple radiation beams converging at a voxel in the tumor;

Fig. 6 is a geometric diagram showing the various spatial variables relevant for the calculation of the individual beam weights;

Fig. 7 is a flow chart showing the steps of the present invention for determining the beam weights for the beams controlled by the collimator of Figs. 1-3 at various stage angles;

Fig. 8 is a schematic representation of a fast electrostatic computer that can be used to determine a charge distribution in a tumor-shaped area as simulated by a number of conductive elements; and

Fig. 9 is a detailed circuit diagram of such a conductive element and its connection to adjacent connecting elements or to a charge amplifier or to a voltage source through a solid-state switch.

Detailed description of the appropriate design

In Fig. 1, a radiation therapy unit 10 suitable for use with the present invention includes a radiation source 12 which produces a generally conical radiation beam 14' radiated from a focal point 18 and directed toward a patient 17 (not shown in Fig. 1). The conical beam 14' is collimated by a radiopaque mask 16 constructed from a set of rectangular collimator vanes to form a generally planar fan beam 14 centered against a fan beam plane 20.

1. The compensator

Before the patient 17 receives the radiation, a compensator 22 is centered in the fan beam 14 and against the fan plane 20, comprising a plurality of adjacent trapezoidal leaves 30 which together form a arc of constant radius around the focal point 18. The blades 30 are held in collars 24. The collars 24 are constructed of radiation-transmissive materials and are attached at their inner ends 23 to a base plate 26 fixed with respect to the focal point 18. The base plate 26 is constructed of a strong, radio-opaque material and is mounted just outside the fan beam 14 to prevent interference with the fan beam 14.

In order to divide the fan beam 14 into a set of adjacent panel-like beams 28 at offset angles φ, the blades 30 of the compensator 22 preferably face the entire fan beam 14. Also in Fig. 2, each sleeve 24 is open at its outer end 27 to slidably receive a trapezoidal blade 30 of comparable size constructed of a dense, radiopaque material such as lead, tungsten, cerium, tantalum, or a related alloy.

Each blade 30 can slide completely within its corresponding sleeve 24 to block the individual jet 28 associated with that sleeve 24. When the blade 30 blocks its corresponding individual jet 28, it is said to be in a "closed state." The sleeves 24 have sufficient length to allow each blade 30 to slide out of the path of the fan jet 14 so as to leave its corresponding individual jet 28 completely unobstructed while still being guided by the sleeve 24. In this non-blocking position, a blade is said to be in the "open state."

By means of a corresponding actuator connected to the blade 30 by a flexible connection 34, each blade 30 can be rapidly displaced between its open and closed states. The actuators 32 have internal pistons (not shown) which can be displaced at high speed between the ends of the actuators 32 by means of compressed air connected to the actuators 32 by supply hoses 35. The supply hoses 35 are fed by a compensator control (not shown in Figs. 1 or 2) to be described below. In order to rapidly and independently displace the blades 30 between the open and closed states, the actuators 32 can exert high forces on them.

In Figures 2 and 3, the blades 30 are supported and guided in the collars 24 by guide rails 36 fitted into grooves 38 cut along the edges of the blades 30. The grooves 38 enable the guide rails 36 to slidably hold the blades 30 in the collars 24 during movement between the open and closed states.

In the closed state, the inner end 40 of each blade 30 is received by a rigid bezel 42 attached to the base plate which aligns the blade 30 with the base plate 26 and hence with the fan beam 14 more precisely than would be possible with the guide rails 36. While the ideally radiolucent guide rails 36 are relatively insubstantial, the bezel 42 mounted on the base plate 26 outside the fan beam 14 need not be radiolucent in contrast and is therefore more substantial in construction. A bezel similar to the bezel 42 (not shown) supports each blade 30 when it is fully in the open state. Since the blades 30 remain in the fully open or fully closed state most of the time, Most of the time they are firmly attached to a support frame 42.

II. Radiotherapy hardware

In Fig. 4, the radiation source 12 is mounted on a stage 44, the latter rotating in the plane 20 of the fan beam about a center of rotation 45 in the patient 17 such that the fan beam 14 can irradiate a slice of the patient 17 from a variety of stage angles A.

The radiation source 12 is controlled by a radiation control module 48, which switches the radiation beam 14 on or off under the control of a computer 51.

In order to separately control the actuators 32 for moving each blade 30 into and out of its respective sleeve 24 and individual jet 28, a compensator control module 52 provides a source of compressed air and valves for blocking that air through the supply hoses 35 (see also Fig. 1). In addition, the compensator control module 52 is connected to the computer 51 to enable program control of the compensator 22 to be described.

In order to produce a tomographic or slice image of the irradiated slice of the patient 17 for planning purposes prior to radiotherapy, a tomographic imaging system 11 using an X-ray source 46 and an opposing sensor matrix 50 may advantageously be mounted on the same stage 44 as the radiation source 12. Alternatively, such tomographic imaging may be carried out on a separate machine, with the slices being scanned in accordance with the Patients are aligned with 17 scale divisions determined by measurement.

A stage control module 54 generates the signals necessary to rotate the stage 44 and therefore change the position of the radiation source 12 and the angle θ of the fan beam 14 for radiation therapy, as well as the computed tomography x-ray source 46 and the sensor array 50 also attached to the stage 44. The stage control module 54 is connected to the computer 51 so that the stage can be rotated under computer control and also to provide a signal to the computer 51 indicative of the stage angle θ to assist in that control.

The control modules for the tomographic imaging system 11 include: an x-ray control module 56 for turning the x-ray source 46 on and off, and a data acquisition system 58 for receiving data from the sensor array 50 to construct a tomographic image. Those skilled in the art will appreciate that a sensor array 50' may also be arranged to receive radiation from the radiation source 12 through the patient 17 to assist in verifying treatment.

To assist in "reconstructing" a tomographic image from such data according to methods well known in the art, an image reconstructor 60, typically including a fast matrix processor or the like, receives the data from the acquisition system 58. As described below, the image reconstructor 60 also communicates with the computer 51 to assist in the fast calculations used in the present invention. The tomographic image enables the verification of the Patient attitude immediately before radiotherapy treatment.

A terminal 62 having a keyboard and a display unit 63 enables an operator to enter programs and data into the computer 51 and enables the radiotherapy unit 10 and the tomographic imaging system 11 to be controlled and to display tomographic images produced by the image reconstruction device 60 on the display 63. A mass storage system 64, which is either a magnetic disk unit or a tape drive, enables data collected by the tomographic imaging system 11 to be stored for later use.

Computer programs for operating the radiotherapy unit 10 are generally stored in the mass storage unit 64 and loaded into the internal memory of the computer 51 for rapid processing during use of the unit 10.

During operation of the radiation therapy unit 10, the compensator control module 52 receives a flux profile from the computer 51 for each stage angle. The flux profile describes the intensity or flux of each beam portion 28 of the fan beam 14 from the radiation source 12 required for that stage angle 9 for a given position of the patient support table (not shown) displaced by the fan beam 14. The flux profiles for each stage angle together form a treatment sinogram for a particular position of the patient table.

The compensator control module 52 rapidly shifts the blades 30 of the compensator 22 for each individual beam 28 between their open and closed states to either fully attenuate each beam 28 or to not attenuate any attenuation. Gradations in the flux of each sub-beam, as required for each flux profile, are obtained for each stage angle by adjusting the relative duration that each blade 30 is in the closed state compared to the relative duration that each blade 30 is in the open position. The ratio between the closed and open states, or "duty cycle," for each blade 30, affects the total energy transmitted by a given blade 30 at each stage angle and thus controls the average flux of each sub-beam 28. The ability to control the average flux at each stage angle allows precise control of the dose delivered by the fan beam 14 to the irradiated volume of the patient 17 using therapy planning techniques described below.

The particle flow profiles of the treatment sinogram are determined by therapy planning software (described below) and stored in the computer 51.

III. Therapy planning software

In a first embodiment, the generation of a treatment sinogram required to obtain the full benefits of the compensator described above is carried out by specially developed software running on the computer 51 and on the reconstruction device 60. Although the treatment planning is carried out in software, it is recognized that the planning can also be implemented in discrete electronic circuitry dedicated to this operation, and such dedicated circuitry can be used to provide even greater speed to this process.

Finally, certain critical steps of the process can be carried out in a second execution on a special electrostatic computer to be described.

5 and 7, the generation of the target treatment sinogram for controlling the compensator 22 begins with the definition of a target dose map 66. The dose map is a set of dose values identified in spatial coordinates or zones in the slice of the patient 17 to be irradiated. A typical target dose map 66, within a dose constraint, assigns a relatively high radiation dose to a tumor tissue region 68 while assigning a second, lower radiation dose to a region 72 containing a radiation-sensitive organ or the like.

The target dose map 66 can be most easily entered by displaying the tomographic view of the patient's 17 slice on the display 63 of the terminal 62 (such that it would appear similar to Figure 5) and manually tracing around the tumor area 68 and the sensitive area 72 using a trackball or similar input device as is common in the art. Standard area-filling computer programs can be used to input a predetermined uniform dose value to each tracked area into the appropriate element in the memory matrix representing the target dose map 66.

The target dose map 66 is stored as a matrix of elements in the memory of the computer 51, each element holding a numerical value and having a spatial coordinate corresponding to a position in a slice of the patient 17 to be treated.

Each element of the dose map 66 defines the target dose at each of the plurality of volume elements 74 ("voxels") in a slice of the patient 17. In Figure 6, each voxel 74 of the patient 17 can be characterized by a vector x defined by a given reference point 76. The target dose at each voxel 74 is D(x). The input of the target dose map as described above is indicated by the process block 90 and produces the matrix Dm(x) in memory.

Typically, the dose at any given voxel 74 in the patient 17 results from the interaction of a number of different beams 28(a) through 28(d) produced at different stage angles θ. However, the contribution of each beam 28 to the dose at any voxel 74 is generally not equal. This is because each beam 28 necessarily passes over a continuous volume along its path intersecting other voxels 74 to which other doses may be associated. For example, the beam 28(d) irradiating the voxel 74 in the tumor 68 also passes through the voxel 74' in the radiation-sensitive organ 72. Therefore, the dose to the voxel 74 is preferably provided by beams other than 28(d). The measure of the relative strength of each beam is generally referred to as the "single beam weight".

The assignment of the dose contribution among different individual beams 28, ie the determination of the individual beam weights, is an extremely complex problem. The present invention recognizes that a solution to this problem is to model the target dose map as the potential energy of a conductive plate. This recognition allows the use of many well-developed methods previously used to solve electrostatic problems.

Once a charge distribution is determined that would produce a potential field equal to the desired dose map 66, the beam weights for any given beam 28 can be determined simply by summing the charges along the path of that beam 28, as will be described.

Once the target dose map 66 has been determined in process block 90 in Fig. 7, it is used to calculate a charge distribution in a volume corresponding to the treatment disk as indicated in process block 92. The particular charge distribution is that which would produce a potential energy field equal to the target dose map 66.

Any stationary electric charge, such as an electron, creates an electrostatic field around itself. That field is generally a measure of the force exerted on a test charge in the region of the stationary electric charge, and is proportional to the inverse of the square of the separating distance, according to Coulomb's law.

The electrostatic field provides the test charge with a well-defined potential energy that is simply proportional to the inverse of the separating distance. For several charges at positions x', the potential energy at a given position x is as follows:

where P(x) is the potential at the spatial coordinate x,

r(x') the charge distribution at other spatial coordinates x' and

x - x' is the distance between the two coordinates x and x'.

The charge distribution calculation indicated by process block 92 therefore simply requires determining a charge distribution ρ(x) such as P(x) = Dm(x). That is:

In a first method of calculating the charge distribution of the process block 92, the electrostatic analogy is evaluated in a special electrostatic computer 91.

In Figures 8 and 9, the patient's 17 disk in this electrostatic computer 91 is represented by a planar matrix of conductive elements 94 arranged in rectangular columns and rows. Each conductive element 94 may be a square copper plate. Each conductive element 94 is connected along the rows and columns to its neighboring conductive elements 94 by four solid state switches 98 which may be closed to form a conductive connection 100 between any two neighboring conductive elements 94 or which may be open to isolate a conductive element 94 from its neighbors. Thus, when all of the solid state switches 98 are closed, the conductive elements 94 act as a single conductive foil.

In addition, each conductive element is connected via a solid state switch 102 to a charge amplifier 104, which switches the charge on any isolated conductive element 94 can measure, or alternatively, as determined by switch 103, connected to a voltage source 106.

The electrostatic computer 91 receives the dose map 66 and, as indicated by process block 92, calculates the charge distribution by performing the following steps.

First, all conductive elements 94 are connected together and grounded to remove any accumulated charge.

Then, the conductive elements 94 are connected to connections 100 in a zone for receiving a uniform dose value, such as the tumor zone 68, by closing the required solid state switches 98. The connected elements 94 then form a conductive foil with an edge that matches that of the tumor 68 in the patient's slice. The voltage source 106 is then connected to one of the connected conductive elements 94, allowing charge to flow into the connected elements 94 until their voltage potential of the elements 94 reaches the level corresponding to the target dose of the dose map 66. Since the conductive elements 94 form a single conductive foil, the potential of each conductive element 94 is the same, thus allowing only uniform doses to be modeled within a given area. However, since a uniform dose is typically desired over the entire tumor area in a slice, this is not overly limiting. If desired, further subzones can be used to approximate non-uniform dose distributions.

As soon as the target potential is reached on the connected conductive elements 94, the charges on the connected elements 94 to various locations as required to create a uniform potential field across those connected elements 94. At this time, the connections 100 are broken by opening any closed solid state switches 98 in the third step, isolating each conductive element 94. Each conductive element 94 is connected in sequence through solid state switches 102 to a charge amplifier 104 which produces an output voltage equal to the charge on that element 94. These output voltages are read by an analog-to-digital converter (not shown) to produce a matrix in the computer 51 indicating the charge distribution ρ(x).

An alternative method of calculating the charge distribution for the process block 92 is to model the charge distribution in the computer 51. In this case, the calculation of the charge distribution is determined by sequential numerical methods.

A first approach uses the Laplace differential equation with the appropriate boundary condition as determined by the outline of the different zones 68 or 72 defining the tumor or the sensitive organ. The Laplace equation is:

²Φ(x) = 0, (3)

where Φ is the electrostatic potential energy, which must be set equal to the dose. Accordingly, the corresponding equation for the dose distribution is:

²Dm(x) = 0. (4)

The solution to the Laplace equation can be obtained numerically. Solution procedures are described by W. H. Press, B. P. Flannery, S. A. Teukolsky and W. T. Vetterling, Numerical Recipes (Cambridge University Press, 1986).

The solution of equation (4) primarily determines the drop in dose outside the tumor area. Once this potential energy distribution outside the tumor is obtained, the corresponding charge distribution ρ(x) is given by the Poisson equation:

²φ(x) = -4ππ(x), (5)

where the charge distribution is confined to a single plane, as is the case here. If the target dose distribution is uniform, the electric field on the surface of the plate is normal to the plate, where the charge distribution is calculated using the simplified relationship:

where E is the electric field and z is the Cartesian coordinate normal to the plate. It is noted that the dose using this method does not generally have to be uniform over the area of any given zone.

In a third approach, the problem of determining the charge distribution can be solved numerically by discretizing equation (1). This method is called the collocation method. In this case, equation (1) becomes:

where Δv is the area of each voxel xi.

This results in n linear equations equating the dose at a voxel xj to a linear sum of charges weighted by 1/r factors at the n discretized points. This system of equations can be solved using standard linear algebra techniques well known to those of ordinary skill in the art.

A complication of this approach is the determination of the effect of a given charge on itself, as occurs when j = i and the denominator of equation (7) tends to zero. One approach to this problem is discussed in Integral Equation Methods in Potential Theory and Electrostatics by M. A. Jaswon and G. T. Symm, published by Academic Press, London, 1977.

In Fig. 7, the individual beam weights are determined after determining the charge distribution by one of the above methods according to process block 110. In process block 110, the charges defined by the charge distribution ρ(x) are integrated (added together) along the individual beam paths to produce the treatment sinogram.

In Fig. 6, for a given single beam 28, the particle flux of that single beam 28 f(s, φ) defined by an angle φ with respect to a horizontal x-axis through the patient 17 and in the stage plane and by a distance s of closest approach of the single beam to the center of rotation 45 is as follows:

f(s, φ) = dpπ(x), (8)

where:

f(s, φ) is the particle flux for any given single beam;

q is a distance variable along the single ray used as the variable of integration.

Thus, these three steps of process blocks 90, 92 and 110 provide all the information required to generate a treatment sinogram for any number of individual beams.

To simplify the integration process, the individual beams of the fan beam can be recombined into sets of parallel beams at different stage angles for a fan beam system in which the beams of radiation 28 diverge at points along the radius of the radiation source around the origin 45.

For the simple case where the dose map 66 contains a single zone 68 with a positive dose that does not vary too rapidly in space, the sinogram of single beam weights at various stage angles θ includes only non-negative single beam weights. Thus, after the process 110, the sinogram contains no negative single beam weights at decision block 112, and the program that generated the necessary treatment sinogram exits at exit point 114.

For multi-zone dose maps 66 or those dose maps that have concave zones or radiation sensitive zones 72 near the tumor zone 68, the steps of process blocks 90, 92 and 110 often provide a sinogram with negative single beam weights. Negative single beam weights represent radiation single beams which do not transmit a dose when passing through the patient 17, but absorb a dose. Such negative individual beam weights cannot be physically realized. If negative individual beam weights are generated, an iteration method is used to eliminate the negative individual beam weights, which nevertheless approximates the target dose characteristic field 66.

At process block 116, the single beam weights of the sinogram are checked, replacing negative single beam weights to produce a truncated sinogram with zero single beam weights, the minimum realizable single beam weight.

This truncated sinogram is used to determine an actual dose that the patient would receive using the individual beam weights of the truncated sinogram. The actual dose D'(r, θ) from all individual beams 28 converging at each voxel 74, defined in standard polar coordinates as shown in Figure 6, is at each voxel 74:

D'(r, θ) = dφf[r sin (φ - θ), φ]. (9)

As indicated by process block 118, equation (9) is simply the superposition of the dose contributed by each individual ray 28 intersecting a given voxel 74 and summing their contributions.

After cutting off the individual beam weights of process block 116, the actual dose D' generated by process block 118 typically differs from the target dose D originally entered in process block 90. Accordingly, an error dose characteristic field ΔD is calculated by subtracting the value of the target dose D from the actual dose D' determined by the process block 118 for each voxel 74. Since negatively weighted individual beams were cut off, it is to be expected that the actual dose D' is usually higher than the target dose D.

Normally, the dose map is updated at each iteration. If further iterations show no changes in the resulting actual dose, the procedure is considered to have converged. This is an indication that the best possible beam distribution has been found.

Typically, the error dose map ΔD is not acceptable in the initial iteration, where the program branches to process block 124 and modifies the dose map Dm entered into the charge distribution calculator of process block 92. In the simplest implementation, the new dose map, referred to as the "current dose map", is equal to the error dose map, (r · ΔD), multiplied by a constant r, which is subtracted from the last dose map provided to process block 92. For the first iteration, the error dose map ΔD is subtracted from the target dose map D, while the target dose map Dm is obtained for subsequent iterations. The modified target dose map is referred to as the current dose map 66' and is provided again to process block 92 for the calculation of a new charge distribution, as indicated by a loop of the program from process block 124 back to process block 92.

For all iterations after the first iteration, the instantaneous dose map 66' (Dm) is corrected by the error dose map ΔD.

The inventors believe that this adjustment of the instantaneous dose map (Dm) creates an iterative solution that rapidly converges to the best treatment sinogram.

Example I

In a computer simulation, a grid of nine tumor target voxels with -1 ≤ x ≤ 1 and -1 ≤ y ≤ 1 was formed. An organ voxel to be avoided had coordinates x = 3, y = 0. The rays were incident with θ values n/2, 7/12π, 11/12π; from each θ value, seven individual rays emerged, each with an angular width of 0.174 radians (about 10º). The described iteration procedure converged in 10 iterations to provide the best possible set of beam weights. The solution provides a dose uniformity in the tumor target to within ±2%, with only a maximum dose of 6% delivered to the organ.

The above description relates to a practical embodiment of the present invention. It will be apparent to those skilled in the art that many modifications can be made. For example, the dose in any given zone need not be uniform, but can be varied as research indicates for the optimal dose distribution in a tumor. Obviously, the method of planning radiotherapy is not limited to a particular radiation source, but can be used with any radiation source that can be broken down into separate beams of radiation.

Claims (11)

1. Radiotherapy device provided with a radiation source (12) for generating a radiation beam (14) which is directed at a volume of a patient at several stage angles lying in a stage plane and contains several individual beams distributed in the stage plane, each individual beam having an independently adjustable particle flux, an individual beam following a path at a single stage angle, and with a therapy planning device for determining the particle flux of each beam, the therapy planning device containing: an electronic computer (51) which in turn contains: means for receiving a target dose map for a target irradiation dose along a slice through the patient's volume; characterized in that the therapy planning device further contains: a charge calculation device for calculating a distribution of electrical charges in the disk that would generate a potential energy field corresponding to the target characteristic field; and an integration device for integrating the charges of the distribution lying along the path to determine the particle flux of the single beam at the stage angle. 2. Radiotherapy device according to claim 1, wherein the charge calculation device comprises a conductive Foil (91) formed from a plurality of connectable conductive elements (94) which are first connected together to receive a charge in at least one dose region and then electrically isolated from each other to measure the charge of each isolated element. 3. Radiotherapy apparatus according to claim 1, wherein the charge calculation means is a part of the computer (51) which receives the dose map as a matrix of numerical values and models the charge distribution according to the following relationship: where: D(x) is the target dose at the spatial coordinate x , ρ(x') is the charge distribution at other spatial coordinates x' and x - x' is the distance between the two coordinates x and x'. 4. Radiotherapy device according to claim 1, wherein the means for receiving a desired dose map has, first, a tumor zone for receiving at least one treatment irradiation threshold value and, second, an organ zone for receiving at most one tolerance irradiation threshold value, the two zones lying along a slice through the volume of the patient, the means being intended to generate an instantaneous dose map equal to the desired dose map; the charge calculation device is designed to receive the current dose map in order to To calculate the distribution of electrical charges in the tumor zone and in the organ zone, which would generate a potential energy field corresponding to the dose map in each zone. 5. A radiotherapy device according to claim 4, further comprising: (a) means for repeating the integration of the charges for the plurality of individual beams at a plurality of stage angles to produce a sinogram of particle flux values; (b) means for setting any negative flux values in the sinogram to a predetermined non-negative value to produce a matched sinogram; (c) means for determining the actual dose that would be produced by individual beams having the flux values of the fitted sinogram to produce an actual dose map; (d) means for evaluating the difference between the actual dose map and the desired dose map in order to modify the current dose map; where the instantaneous dose map can be repeatedly adjusted to produce an actual dose map that closely approximates the target dose map and does not contain any individual beams with negative particle flux values. 6. Radiotherapy device according to claim 5, wherein the current dose map is modified by subtracting the difference between the actual dose map and the target dose map from the current dose map. 7. A radiotherapy device according to claim 5, wherein the charge calculation means determines the distribution of the charges according to the following relationship: where: D(x) is the target dose at the spatial coordinate x , ρ(x') is the charge distribution at other spatial coordinates x' and x - x' is the distance between the two coordinates x and x'. 8. Radiotherapy device according to claim 1 or 5, wherein the integration device determines the particle flux of the single beam along the path according to the following relationship: f(s, φ) = dqρ(x, y) where f(s, φ) is the particle flux for any given single beam; q is a distance variable along the particle beam that is used as an integration variable ; and ρ(x, y) is the charge distribution over the disk. 9. Method for radiation therapy planning for a radiation therapy device with a radiation source (12) for generating a radiation beam (14) which is directed at a patient's disc at several stage angles lying in a stage plane and contains several individual beams distributed in the stage plane, each individual beam having an independently adjustable particle flux, wherein an individual beam follows a path through the disc at a stage angle, the method comprising: (a) entering a target dose map along a slice through the patient's volume as set digital values into an electronic computer (51), the digital values representing a dose of ionizing radiation necessary for treating the patient; characterized in that the method further comprises : (b) calculating a distribution of electrical charges in the disk that would produce a potential energy field corresponding to the target dose map; and (c) Integrate the charges of the distribution lying along the path to determine the particle flux of the single beam at the stage angle. 10. The method of claim 9, wherein the target dose map has firstly a tumor zone for receiving at least one radiation treatment threshold and secondly an organ zone for receiving at most one radiation tolerance threshold, the two zones lying along a slice through the volume of the patient to produce an instantaneous dose map equal to the target dose map; (b) the calculation of the distribution of the electrical charges calculates those charges within the tumor zone and the organ zone that would generate a potential energy field corresponding to the current dose map in each zone. 11. The method of claim 10, further comprising: (c) repeating the integration of charges for the plurality of individual beams at the plurality of stage angles to produce a sinogram of particle flux values; (d) setting any negative flux values in the sinogram to a predetermined non-negative value to produce a fitted sinogram; (e) determining the actual dose that would be produced by the individual beams with the particle flux values of the fitted sinogram to produce an actual dose map; (f) means for evaluating the difference between the actual dose map and the target dose map in order to modify the current dose map; (g) Repeat steps (b) to (f) until further repetitions do not change the difference.